Identification Of Dynamic Load Location Based On Genetic Algorithm

Dynamic load identification technology occupies an important loaction in engineering as a discipline developed based on practical problems and always been a difficult problem in dynamics.In order to avoid the ill-posed problem of matrix inversion in dynamic load identification,this paper combines genetic algorithm to study the identification of dynamic load loaction.The positive time domain model of dynamic load loaction identification is proposed,combining the system dynamics equations and optimization theory.On this basis,the optimization identification method combined with genetic algorithm was studied.And decision variables,objective functions and constraints are determined.Furthermore,the recognition algorithm is designed,and the setting and selection of key parameters are analyzed.Based on the above theory,two identification methods of parameterization and time series are proposed.Based on Fourier series expansion and Chebyshev polynomial theory,the parameterized method determines the parameter expression form of the dynamic load and the identification model.The time series method combines the discrete time system idea and the step force assumption to establish a discrete recognition model.According to the above different recognition models,the key parameters and process of the recognition algorithm are determined.Computer simulation examples verify that the above two methods are feasible and effective.For complex engineering structures,a step-by-step identification method combined with finite element is proposed.Combined with the idea of fast approximation,this method reduces the workload of model calibration and optimization.The simulation shows that the method is effective.Finally,the experimental verification of dynamic load location identification was carried out.The measured model parameters and dynamic response data were obtained through vibration experiments.Then,through the forward optimization identification method,the loaction of the experimental dynamic load is identified.The recognition result further proves that the method in this paper is effective and has the feasibility of practical application.